ExerciseIncreasesInsulinContentandBasalSecretionin ......improvements in exercise-induced glycemic control in T1D mice may be due to enhancement of insulin content and secretion in

Hindawi Publishing CorporationExperimental Diabetes ResearchVolume 2011, Article ID 481427, 10 pagesdoi:10.1155/2011/481427

Research Article

Exercise Increases Insulin Content and Basal Secretion inPancreatic Islets in Type 1 Diabetic Mice

Han-Hung Huang,1 Kevin Farmer,1 Jill Windscheffel,1 Katie Yost,1 Mary Power,1

Douglas E. Wright,2 and Lisa Stehno-Bittel1

1 Department of Physical Therapy and Rehabilitation Science, University of Kansas Medical Center, 3901 Rainbow Blvd.,Kansas City, KS 66160, USA

2 Department of Anatomy and Cell Biology, University of Kansas Medical Center, 3901 Rainbow Blvd.,Kansas City, KS 66160, USA

Correspondence should be addressed to Lisa Stehno-Bittel, [email protected]

Received 19 April 2011; Accepted 17 May 2011

Academic Editor: N. Cameron

Copyright © 2011 Han-Hung Huang et al. This is an open access article distributed under the Creative Commons AttributionLicense, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properlycited.

Exercise appears to improve glycemic control for people with type 1 diabetes (T1D). However, the mechanism responsible forthis improvement is unknown. We hypothesized that exercise has a direct effect on the insulin-producing islets. Eight-week-oldmice were divided into four groups: sedentary diabetic, exercised diabetic, sedentary control, and exercised control. The exercisedgroups participated in voluntary wheel running for 6 weeks. When compared to the control groups, the islet density, islet diameter,and β-cell proportion per islet were significantly lower in both sedentary and exercised diabetic groups and these alterationswere not improved with exercise. The total insulin content and insulin secretion were significantly lower in sedentary diabeticscompared to controls. Exercise significantly improved insulin content and insulin secretion in islets in basal conditions. Thus, someimprovements in exercise-induced glycemic control in T1D mice may be due to enhancement of insulin content and secretion inislets.

1. Introduction

Across the globe, the incidence of type 1 diabetes (T1D) isincreasing at an alarming rate with the age of onset decreas-ing within a 20-year period [1]. T1D is an autoimmunedisease that leads to impaired glucose homeostasis, becausethe insulin-producing cells (β-cells) located in the pancreaticislets of Langerhans are attacked by the immune system [2].The clinical benefits attributed to exercise for individualswith type 2 diabetes have been well documented [3–7]. InT1D, far fewer reports have focused on the role of exercise,and yet there are strong indications for decreased riskof diabetes-associated complications including improvedarterial function [8–11], cardiac performance [12–17], andlipid metabolism [18] resulting in lower total cholesterol,LDL-c, and triglycerides [19, 20].

Only recently have studies focused on the relationshipbetween exercise and blood glucose regulation in T1D. In

general, these reports have shown reduced insulin dosagewith increased physical activity [21] along with better gly-cemic control [22]. In fact, moderate to vigorous activity hasbeen associated with greater overall fitness, a higher fat freemass, and lower glycosylated hemoglobin (HbA1c) levels inpeople with T1D [19, 20]. In a randomized study of 196adults with T1D, those that exercised moderately once tothree times per week significantly reduced the HbA1c levelsand insulin requirements [23]. Even more convincing is alarge cross-sectional study of over 19,000 children with T1D,finding that the amount of physical activity was one of thestrongest factors predicting lower HbA1c values [24].

In spite of these impressive results, little is known aboutthe mechanism of action to explain the lowered bloodglucose levels with exercise. There are two general sites whereexercise could directly affect blood glucose regulation: (1)insulin secretion of the islets and (2) insulin-stimulatedglucose uptake in the skeletal muscle [18, 25, 26]. Only

mailto:[email protected]

2 Experimental Diabetes Research

two papers have directly tested the first possibility: changesin insulin levels with exercise in T1D. Unfortunately, theonly assay used was an immunohistochemical measurementof the number of hormone-positive cells in the islets ofexercised diabetic rats [27, 28]. Clearly, more must be doneto begin to unravel the cellular changes occurring in isletswith exercise training. We utilized a T1D animal model toexamine islet morphology, density, size, cell composition,insulin secretion, and insulin content after 6 weeks ofvoluntary exercise. By testing the hypothesis in an extremelysevere form of T1D without insulin treatment, we ensuredthat any changes we identified in the islets were due directlyto the effects of exercise and not secondarily to a reduction ofblood glucose levels.

2. Materials and Methods

2.1. Animals. The animal protocols were approved by theInstitutional Animal Care and Use Committee. Fifty-three 8-week-old male A/J mice were divided into four groups: (1)sedentary (nonexercised) diabetic (n = 18), (2) exerciseddiabetic (n = 13), (3) sedentary control (n = 11), andexercised control (n = 11). Both diabetic groups wereinjected with streptozotocin, toxic to islet β-cells, (freshlydissolved in 10 mM sodium citrate, pH 4.5, with 0.9%NaCl, (Sigma, St. Louis, Mo, USA)) on consecutive days(1st injection of 85 mg/kg body weight and 2nd injection of65 mg/kg body weight) to induce diabetes. Control animalswere injected with 0.4 mL of sodium citrate buffer, pH4.5. Mice were fasted 3 hours before and 3 hours aftereach injection. Blood glucose levels were monitored weekly(glucose diagnostic reagents; Sigma) after fasting for 2 hours.Mice in the two diabetic groups were identified as havingdiabetes when blood glucose levels reached greater than190 mg/dL.

2.2. Exercise. After the induction of diabetes, the mice fromthe exercise groups were individually housed in cages con-taining exercise wheels (Mini Mitter Co. Inc., a RespironicsCompany, Bend, Ore, USA) for 6 weeks starting 3 days afterthe 2nd injection of streptozotocin. Each wheel revolutionwas continuously recorded and summarized in thirty-minuteintervals with the Vital View Data Acquisition System (MiniMitter Co. Inc.) throughout the duration of the study. At thetermination of the study, mice were overanesthetized withavertin after a 2-hour fast and the pancreata either processedfor immune-staining or for isolated islet retrieval.

2.3. Pancreatic Sections. The pancreata were rapidly har-vested and fixed in 10% normal buffered formalin. The tis-sues were subsequently dehydrated in graded concentrationsof ethanol, cleared in xylene, and subsequently embeddedin paraffin wax at 55◦C. The tissues were sectioned at8 μm thickness, mounted on Superfrost/Plus microscopeslides (Fisher Scientific, Pittsburg, Pa, USA, no. 12-550-15) and dried at 40◦C overnight and stored at 4◦C untilprocessing. The paraffin-embedded sections were deparaf-finized/rehydrated in xylene followed by ethanol and PBS

serial rehydration. Antigen retrieval was completed in 0.01 Mcitrate buffer, pH 6.2, with 0.002 M EDTA for 30 min usinga steamer. Cells were permeabilized in 1.0% Triton X-100 in0.1 M PBS, pH 7.4 for 30 min. Hematoxylin and eosin (H&E)staining was performed on some sections to illustrate thegeneral islet morphology under light microscopy.

2.4. Immunofluorescence Staining. Sections were blocked in10% normal donkey serum, 1.0% bovine serum albumin(BSA), and 0.03% Triton X-100 diluted in 0.1 M PBS, pH7.4 for 30 min. Incubation with the primary antibody mixwas performed at 4◦C overnight in a wet chamber followedby incubation with the mix of fluorophore-conjugatedsecondary antibody at room temperature for 2 hr in a wetchamber protected from light. Both primary and secondaryantibodies were diluted in 1% NDS, 1% BSA, and 0.03%Triton X-100. Slides were mounted with antifading the agentGel/Mount (Biomeda, Foster City, Calif, USA, no. M01). Thefollowing primary antibodies were used: anti-insulin (1 : 100,Abcam, Cambridge, Mass,USA, no. ab7842), antiglucagon(1 : 200, Abcam, no. ab10988), and antisomatostatin (1 : 200,Abcam, no. ab53165). Corresponding secondary antibodieswere conjugated with Cy2 (1 : 200, Jackson ImmunoResearchLaboratories Inc., West Grove, Pa,USA, no. 706-225-148),Alexa 555 (1 : 400, Molecular Probes, Eugene Ore,USA, no.A31570), and Alexa 647 (1 : 400, Molecular Probes, no.A31573). Images were collected using a Nikon C1si confocalmicroscope.

2.5. Islet Density, Size, and Cell Composition. Islet densitywas defined as the numbers of islets per microscopic field.Four random fields were selected per section. Twenty-fournonserial sections of pancreatic tissues from 3 animals pergroup were evaluated. Islet size was estimated by measuringthe diameter of the islet under the software Nikon EZ-C13.0 FreeViewer. For cell composition analysis, the relativeproportion of α-,β-, or δ-cells in each islet was evaluated bycounting the number of individual types of cell and dividingby the total sum of endocrine (α, β, and δ) cells per islet.For each group, 88 to 106 islets from nonserial sections ofpancreatic tissues from 3 animals were analyzed for bothcell size and cell composition. All the evaluations were doneby three blinded examiners with a high interrater reliability(intraclass correlation coefficient; ICC > 0.8).

2.6. Insulin Content In Situ. To determine the insulin contentin the pancreatic islets in situ, insulin immunohistochemistry(IHC) was performed on paraffin-embedded pancreatic tis-sue (Histostain-Plus, Invitrogen) using anti-insulin antibody(1 : 100, Santa Cruz Biotechnology, Inc., no. sc-9168). Afterstaining, slides were dehydrated in xylene and placed on cov-erslips in Permount mounting medium (Fisher Scientific, no.S15-100). The specificity of insulin immunoreactivity wasconfirmed by omitting the primary antibodies from somesections. Images were analyzed with Ps Adobe PhotoshopCZ4 extended software by determining the average pixelvalue of staining per islet and per single cell, using our pub-lished procedure [29]. Background staining was subtracted

Experimental Diabetes Research 3

from each value. Eighteen to 25 islets were randomly selectedfrom nonserial sections of pancreatic tissues from 3 to 5animals per group.

2.7. Islet Isolation. Islet isolation methods followed our pub-lished procedures described in detail [30–32]. Briefly, micewere anesthetized with an intraperitoneal injection of avertin(20 mg/kg). After the peritoneal cavity was exposed, thepancreatic main duct to the intestine was clamped and thepancreas cannulated in situ via the common bile duct. Thepancreas was distended with collagenase and removed. Isletswere gently tumbled, washed, and passed through a sterile30-mesh stainless steel screen and centrifuged. The pellet wasmixed with histopaque, centrifuged, and the islets floatingon the gradient were collected and sedimented. Islets werepassed through a sterile 40 μm mesh cell strainer with Hank’sbuffered salt solution (HBSS). After this cleaning process,islets were placed into CMRL 1066 medium containing 2 mMglutamine, 10% fetal bovine serum (FBS), and 1% antibiotic/antimycotic solution and put into a 37◦C culture chambercontaining 5% CO2.

2.8. Insulin Secretion and Content. Glucose-induced insulinstatic secretion assays followed our published procedures[31]. Isolated islets were placed in 24-well plate with morethan 10 islets per well and assigned to two groups: lowglucose (3 mM) and high glucose (16.6 mM). All wells werepreincubated for 2.5 hours in RPMI 1640 containing 10%fetal bovine serum and 3 mM glucose in a 37◦C containing5% CO2. After preincubation, media was removed from eachwell and discarded. Low (3 mM as basal condition) or high(16.6 mM) glucose solutions were added, according to thedesign. After 30 minutes static incubation in the 37◦C and5% CO2, the islets were sedimented and the conditionedmedium was collected and frozen at −80◦C. At the end ofthe static incubation, the islets were harvested and frozenat −80◦C. The total protein in the islets was extracted bysonication in acid ethanol (0.18 M HCl in 95% ethanol). Thereleased insulin and the total intracellular insulin amountswere determined by the ELISA (ALPCO, Salem, NH, USA)as we have published previously [31]. Insulin secretion andcontent was normalized to islet numbers using previouslypublished standard procedures [33–36] and by volume (isletequivalents) [30–32]. The experiment was replicated 4-5times per group.

2.9. Statistics. Results were expressed as averages of eachgroup or cell population ± SEM and were compared usingANOVA with Fisher least significant difference, repeatedmeasures ANOVA, or student’s t-test, depending on theexperimental design. Significant differences were defined asP < 0.05.

3. Results

The exercised diabetic and exercised control mice were givenunlimited access to voluntary running wheels for 6 weeks.The average running distance was 2.9 ± 0.3 km/day for the

exercised diabetic group and 4.1 ± 0.4 km/day for the exer-cised control group (P < 0.05). In the dependent variables(body weight, fasting glucose levels, islet morphology, insulincontent, and insulin secretion), there was no difference in theparameters between the two nondiabetic groups (sedentarycontrol and exercised control). Since the focus of the studywas the effects of exercise on islets and no differences werenoted between the two control nondiabetic groups, wecompared the data from the two diabetic groups (sedentarynonexercised, diabetic mice, and exercised diabetic mice) tothe sedentary, nonexercised controls.

Over the 6-week study, the sedentary control groupexhibited a 13% increase in body weight. Both diabeticgroups exhibited significant reductions in body weightcompared to controls (P < 0.001). The sedentary and exer-cised diabetic groups displayed a 14% and 11% reductionin body weight over the 6-week experiment, respectively(Figure 1(a)).

Fasting glucose levels were significantly higher in bothdiabetic groups (exercised and sedentary) compared tothe controls (P < 0.001). Among the diabetic groups,blood glucose levels trended upward over the course ofthe study to levels above 300 mg/dL. When compared tothe sedentary diabetic group, the exercised diabetic groupshowed significantly lower glucose levels at weeks 1 and 2(P < 0.05) (Figure 1(b)). However, the difference was notstatistically significant at later time points.

3.1. Islet Morphology. Cellular atrophy and extensive vac-uoles were present in 80% of the islets from sedentarydiabetic mice (Figure 2). No overt vacuoles were noted inthe tissue samples from the control, nondiabetic animals.The same characteristics (cellular atrophy and vacuoles) wereidentified in 73% of the islets from the exercised diabeticgroup. Thus, among diabetic animals, voluntary exercise hadno significant effect on the proportion of islets with vacuoles.

3.2. Islet Density and Size. Diabetes negatively affected thenumber of islets per area, as the islet density was significantlylower in the two diabetic groups compared to the control(P < 0.05). Typical islet density from a control animal isshown in the pancreatic section triple stained for insulin,glucagon, and somatostatin (Figure 3(a)). Figure 3(b) sum-marizes the islet density for sedentary control, diabetic,and exercise-trained diabetic animals. The sedentary diabeticand exercised diabetic mice had significantly lower isletdensities than controls (33% and 30%, resp.). There wasno statistically significant difference between two diabeticgroups.

Diabetes caused a significant decrease in the diameterof the individual islets in both the exercised and sedentarygroups (Figure 3(c)). The sedentary diabetic group hada 23% smaller islet size (diameter) compared to controlanimals, while the exercised diabetic group had 28% smallerislets compared to the nondiabetic controls (P < 0.05).

3.3. Islet Cell Composition. Pancreatic islets in mammalsare composed of several types of endocrine cells, primarilyinsulin-producing β-cells, glucagon-producing α-cells, and


Week

Bod

yw

eigh

t(g

)

16

18

20

22

24

26

ControlDiabetesDiabetes + exercise

10 2 3 4 5 6

∗∗ ∗∗ ∗∗∗∗

∗∗ ∗∗

(a)


Week

1 2 3 4 5 6

Glu

cose

leve

l(m

g/dL

)

0

100

200

300

400

500

∗∗

(b)

Figure 1: Body weight and glucose levels. (a) Body weight was measured weekly showing a decline in weight with diabetes. ∗∗Indicatessignificant differences (P < 0.001) in control (circles) versus the two diabetic groups (squares and triangles). (b) Both diabetic groups hadsignificantly higher blood glucose levels during each week, when compared to the nondiabetic sedentary mice (circles). Blood glucose for theexercised diabetic group (triangles) was lower than the sedentary diabetic group (squares) during the first two weeks. ∗Indicates significantdifference (P < 0.05) in nonexercised diabetic versus exercised diabetic groups. Data from all animals in each group were analyzed for thefigure at each time point; control (n = 11) and sedentary diabetic (n = 18) and exercised diabetic (n = 13).

(a) Control (b) Diabetes (c) Diabetes + exercise

Figure 2: Cellular atrophy and appearance of vacuoles with diabetes. Diabetes was associated with cellular atrophy and increased numbers ofvacuoles in the tissue when compared to healthy islets from control animals (a). There were no significant characteristic differences betweenthe sedentary diabetic and the exercised diabetic groups. (Scale bar: 50 μm.)

somatostatin-producing δ-cells. Immunofluorescent stain-ing was used to determine the cell composition of individualislets (Figure 4(a)). In islets from the control group, β-cellscomprised the major (80%) cell type. However, in the isletsfrom the diabetic animals, the alpha cells were the majority(56%) cell type (Figure 4(a)). The percentage of insulin-positive, glucagon-positive, and somatostatin-positive cellsfrom each group of animals is shown in Figure 4(b).The percentage of insulin-positive cells was significantly(P < 0.05) lower in islets from sedentary diabetic rats

(20.0 ± 1.4%) as compared to the controls (80.4 ± 1.0%).The percentage of glucagon-positive cells was significantlyhigher (P < 0.05) in islets from sedentary diabetic rats(56.0 ± 2.0%) as compared to the controls (9.2 ± 0.7%).Finally, the percentage of somatostatin-positive cells was alsosignificantly higher (P < 0.05) in islets from sedentarydiabetic rats (24.0 ± 1.2%) as compared to the controls(10.4 ± 0.8%). The cell composition was nearly identicalbetween the sedentary and exercise-trained diabetic animals(NS, Figure 4(b)).


(a)

0

0.2

0.4

0.6

0.8

1

1.2

1.4

Control Diabetes Diabetes+

exercise

∗

Isle

t n

umbe

r/m

m2

(b)

∗


exercise

Dia

met

er(μ

m)

0

20

40

60

80

100

120

140

(c)

Figure 3: Islet density and size. (a) Four islets are shown in this 10x field of pancreatic tissue from a healthy control animal. (Green: insulin-positive β-cells; red: glucagon-positive α-cells; blue: somatostatin-positive δ-cells.) (Scale bar: 200 μm.) (b) The islet density was determinedas the islet numbers normalized per the area of pancreatic section. Exercise did not increase the number of islets per area in the diabeticgroups. (c) Exercise did not increase the islet’s diameter in the diabetic groups. ∗Indicates significant differences (P < 0.05) between controland the two diabetic groups.

3.4. Insulin Content. Diabetes significantly reduced the totalinsulin content of isolated islets from the three groups ofanimals. Insulin content was normalized to islet using pre-viously published procedures [34, 36]. The average insulincontent in the control group was 62.0 ± 4.9 ng/islet. Forthe sedentary diabetic rats, the insulin content was 0.4 ±0.1 ng/islet. Exercise increased the insulin content by morethan 3 times over the sedentary diabetic animals (1.5 ±0.4 ng/islet). Figure 5(a) compares only the two diabeticgroups as the control value was too large to effectively fit onthe scale.

In addition to the in vitro measurements, we analyzedinsulin content in situ using published procedures [28, 29].First, insulin immune reactivity per islet was calculated usingpublished techniques [37]. In the islets from control animals,there was strong insulin staining in all β-cells (Figure 5(b)).However, in the sedentary diabetic groups, there was a weak

insulin immunoreactivity and it was found in only a few β-cells within the islets. Interestingly, in the exercised diabeticgroup, there was an increase in the insulin labeling intensity(Figure 5(c)). As shown in Figure 5(c), the intensity ofinsulin-positive staining per islet in the exercised diabeticgroup was significantly higher compared to the sedentarydiabetic group (P < 0.001). Exercise improved the totalinsulin content per islet from 48% to 72% when normalizedto the control levels. Further analysis of the insulin intensityper β-cell also showed significantly higher insulin intensityin the exercised diabetic group compared to the sedentarydiabetic group. Exercise improved the individual cell insulinlabel intensity from 92% to 114% when normalized to thecontrols (Figure 5(d); P < 0.001).

3.5. Insulin Secretion. The glucose stimulated insulin secre-tion was measured in low (3 mM) or high (16.6 mM) glucose


(a)

0

20

40

60

80

100


exercise

Cel

lcom

posi

tion

(%)

∗ ∗# #

NS

β-cellsα-cells

δ-cells

(b)

Figure 4: Islet cell composition. (a) Representative islets from each group are shown. In the healthy control (left image), insulin-positive β-cells (green) are the most common cell type. However, in both diabetic groups, [diabetic (center image) and diabetic + exercise (left image)]glucagon-positive alpha cells were the major cell type. (Scale bar: 100 μm.) (b) Graph showing the relative proportion of three major typesof endocrine cells in the islets from each group. There was no significant difference in the cell composition between the sedentary diabeticand exercised diabetic groups (NS). ∗Indicates significant difference (P < 0.05) in β-cells (green) versus α-cells (red) and δ-cells (blue) inthe control. #Indicates significant difference (P < 0.05) in β-cells (green) between control and diabetic groups.

conditions for 30 minutes. Under low glucose, the insulinsecretion in the two diabetic groups was significantlylower than the control group (Figure 6(a), P < 0.001).Interestingly, the insulin secretion in the exercised diabeticgroup was significantly (P < 0.05) higher when comparedto the sedentary diabetic group (Figure 6(a)). Under highglucose, the insulin secretion in the diabetic groups was againsignificantly lower than in the control (P < 0.001). However,there was no significant difference between the exerciseddiabetic group and sedentary diabetic group. These resultswere the same whether insulin secretion was normalized byislet or by volume (islet equivalent).

When the insulin release was normalized for the totalinsulin content, exercise increased the percent of total insulinreleased in basal conditions by 2.5 times (Figure 6(b)). Thepercent of released insulin in the islets from sedentarydiabetic mice was 0.38%, and from exercised diabetic mice itwas 0.94% (P < 0.02). With high glucose exposure, exercise

induced a change in the insulin secretion from 0.69 to 0.98%of the total insulin content, although this change did notreach statistical significance.

4. Discussion

In the 1950s, Joslin first suggested that exercise should bean essential component to regulate blood glucose levels ofpeople with T1D, along with a restricted diet and insulintherapy [38]. Yet, today the mechanisms by which exerciseregulates blood glucose in conditions of T1D are unclear.In the 1980s, classic work by Reaven and Chang showedthat exercised rats with T1D had lower plasma glucose andtriglyceride levels than their sedentary counterparts [39].They hypothesized that the results were due to improvedperipheral insulin sensitivity; however, no direct measure ofislet mass or insulin content was reported. The current study


0

0.5

1

1.5

2

2.5

Diabetes Diabetes+

exercise

Insu

linco

nte

nt

/isl

et(n

g) ∗

(a)

Control Diabetes Diabetes + exercise

(b)

∗∗

Insu

lin/i

slet

(pix

elin

ten

sity

)

0

20

40

60

80


exercise

(c)

0

20

40

60

80

100

120

140∗∗


exercise

Insu

lin/β

-cel

l(pi

xeli

nte

nsi

ty)

(d)

Figure 5: Insulin content per islet. (a) Exercise significantly improved the insulin content per islet between the two diabetic groups (P <0.05). (b) Immunohistochemical staining showing the intensity of insulin-immunoreactive (red) cells in islets from pancreatic sections.(Scale bar: 50 μm.) (c) Quantification of the intensity of insulin-positive staining per islet in pancreatic sections. Exercise improved the totalinsulin staining intensity per islet significantly (∗∗P < 0.001). (d) The intensity of insulin-positive staining was calculated per β-cell, andexercise significantly improved the insulin content per cell (∗∗P < 0.001). Control N = 76; sedentary diabetic N = 43; exercised diabeticN = 56.

is the first to measure exercise-induced increases in insulincontent and secretion in isolated islets from diabetic animals.

We hypothesized that exercise could act by protectingislets during the onset of the disease, because in animalmodels of type 2 diabetes, long-term aerobic exercise resulted

in increased islet β-cell proliferation, increased β-cell mass,and a partial sparing of the abnormal islet morphologynoted in the sedentary diabetic rats [40]. Exercise blockedthe age-associated morphological changes in the pancreas,including multilobulated, fibrotic islets [41]. In addition,



High glucose

∗

∗

Low glucose

Insu

lin S

ecre

tion

(n

g/IE

)

0

0.05

0.1

0.15

0.2

0.25

#

(a)

∗∗


High glucoseLow glucose

Insu

linse

cret

ion

(%of

tota

lin

sulin

)

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

#

(b)

Figure 6: Insulin secretion per islet. (a) Insulin secreted from isolated islets in low (3 mM) or high (16.6 mM) glucose was collected for 30minutes. Under low glucose conditions, exercise improved the insulin secretion per islet significantly (#P < 0.05). Under high glucose, therewas no effect of exercise. ∗P < 0.05 control versus both diabetic groups. (b) When the insulin release was normalized for the total insulincontent, exercise increased the percent of insulin released in basal conditions (#P < 0.02). With high glucose exposure, exercise induced achange in the insulin secretion, although it did not reach statistical significance.

aerobic exercise decreased the presence of proinflammatorycytokines in islet cells [42], but did not change the islet geneexpression pattern in type 2 Zucker diabetic fatty rats [43].

Morphological examination of the islets from ourexercised-trained mice failed to demonstrate differencescompared to the sedentary diabetic group. Our findingsare similar to a previous report investigating the effect ofthe exercise on the distribution of α-, β-, and δ-cells andpancreatic polypeptide cells in the islets of streptozotocin-induced diabetic rats [27]. Conversely, another study focusedon β-cell health and exercise in T1D concluded that exercisepartially spared the β-cells from diabetes [28]. A differencebetween the exercise protocols may explain the discordantresults. In our protocol and the previous paper by Howarthet al., which showed no change in β-cell numbers, theexercise protocol was initiated after the induction of diabetes[27]. In the Coskun et al. study, the aerobic exercise protocolwas initiated four weeks prior to the induction of diabetesand the exercise continued for another eight weeks to thetermination of the experiment [28]. Thus, exercise may beable to protect β-cells if initiated prior to the onset of thedisease but has limited or no ability to rescue the β-cells oncelost.

Even though voluntary exercise did not improve the pro-portion of insulin-producing β-cells in the islets of diabeticanimals, it did improve the insulin content in isolated islets.This effect was not an artifact of the isolation procedure,as exercise nearly doubled the total insulin content per isletwhen examined in situ. Our findings are in agreement witha previous publication that showed a beneficial effect on

insulin content in diabetic animals after exercise [28]. Theprevious publication relied solely on insulin immunohis-tochemistry to draw conclusions. The data presented hereconfirm that finding, but also demonstrate that exercisewas capable of improving the total insulin content andinsulin secretion, when measured directly with quantifiablemethods.

One important difference must be highlighted betweenthe previous studies and the current work. In the Coskun etal. paper, the serum glucose measurements were statisticallylower in the exercising animals than the sedentary diabeticrats [28]. Thus, one cannot rule out the possibility that anyimprovements in β-cell numbers or function were due to thelowered blood glucose values. In the present study, bloodglucose levels were not significantly different between theexercised and sedentary diabetic rats at the termination ofthe study. This was accomplished by inducing severe diabeteswithout insulin treatment, and it provided an advantagewhen interpreting the data, because any changes were directlyrelated to the effects of exercise without a reduction inblood glucose. Since it is known that high glucose is toxicto islet cells [44, 45], maintaining statistically similar bloodglucose readings at the termination of the study ensured thatglucotoxicity was not a factor.

Our central finding is that exercise has an effect on thepancreatic islets by stimulating insulin production and/orsecretion in diabetic animals. The current results clarifythe effect of exercise by showing that the only statisticallysignificant difference in insulin secretion was under lowglucose conditions, whether calculated as insulin secretionper islet, per volume (IE), or per total insulin content. While


this could be important in maintaining normal blood glucoselevels, it would indicate that exercise may not help with post-prandial glucose excursions.

5. Conclusion

In conclusion, this is the first time that insulin contentand secretion has been directly measured in T1D animalsfollowing exercise training. In fact, only two other publica-tions have focused on the direct effects of exercise on isletsunder diabetic conditions [27, 28]. In our study, six weeks ofvoluntary exercise did not alter islet density, morphology, sizeor cellular composition. However, insulin content increasedin situ. Importantly, this is the first study to demonstratethat exercise increased insulin content and improved insulinsecretion in isolated islets in uncontrolled T1D.

Acknowledgments

This work was supported by a grant from National Insti-tutes of Health (NIH R01NS43314 to DW), Juvenile Dia-betes Research Foundation (DW), Emilie Rosebud DiabetesResearch Foundation (LSB), and University of Kansas Med-ical Center Biomedical Research Training Program (HHHand KF). Core facilities and services were supported byNational Institute of Child Health and Human DevelopmentGrant number HD02528.

References

[1] A.-G. Ziegler, M. Pflueger, C. Winkler et al., “Acceleratedprogression from islet autoimmunity to diabetes is causingthe escalating incidence of type 1 diabetes in young children,”Journal of Autoimmunity, vol. 37, no. 1, pp. 3–7, 2011.

[2] A. L. Notkins, “Immunologic and genetic factors in type 1diabetes,” Journal of Biological Chemistry, vol. 277, no. 46, pp.43545–43548, 2002.

[3] S. Bweir, M. Al-Jarrah, A. Almalty et al., “Resistance exercisetraining lowers HbA1c more than aerobic training in adultswith type 2 diabetes,” Diabetology & Metabolic Syndrome, vol.10, p. 27, 2009.

[4] N. LeBrasseur, K. Walsh, and Z. Arany, “Metabolic benefitsof resistance training and fast glycolytic skeletal muscle,”American Journal of Physiology, vol. 300, pp. E3–E10, 2011.

[5] A. Swanson, K. Watrin, and L. Wilder, “Clinical Inquiries: howcan we keep impaired glucose tolerance and impaired fastingglucose from progressing to diabetes?” The Journal of FamilyPractice, vol. 59, pp. 532–533, 2010.

[6] L. VanHoose, Y. Sawers, R. Loganathan et al., “Electrocardio-graphic changes with the onset of diabetes and the impact ofaerobic exercise training in the Zucker Diabetic Fatty (ZDF)rat,” Cardiovascular Diabetology, vol. 9, article 56, 2010.

[7] C. Sanz, J.-F. Gautier, and H. Hanaire, “Physical exercise forthe prevention and treatment of type 2 diabetes,” Diabetes andMetabolism, vol. 36, no. 5, pp. 346–351, 2010.

[8] J. P. H. Seeger, D. H. J. Thijssen, K. Noordam, M. E. C. Cranen,M. T. E. Hopman, and M. W. G. Nijhuis-Van Der Sanden,“Exercise training improves physical fitness and vascularfunction in children with type 1 diabetes,” Diabetes, Obesityand Metabolism, vol. 13, no. 4, pp. 382–384, 2011.

[9] W. G. Mayhan, D. M. Arrick, K. P. Patel, and H. Sun, “Exercisetraining normalizes impaired NOS-dependent responses ofcerebral arterioles in type 1 diabetic rats,” American Journalof Physiology, vol. 300, no. 3, pp. H1013–H1020, 2011.

[10] B. Trigona, Y. Aggoun, A. Maggio et al., “Preclinical Noninva-sive Markers of Atherosclerosis in Children and Adolescentswith Type 1 Diabetes Are Influenced by Physical Activity,”Journal of Pediatrics, vol. 157, pp. 533–539, 2010.

[11] D. M. Roche, S. Edmunds, T. Cable, M. Didi, and G. Stratton,“Skin microvascular reactivity in children and adolescentswith type 1 diabetes in relation to levels of physical activityand aerobic fitness,” Pediatric Exercise Science, vol. 20, no. 4,pp. 426–438, 2008.

[12] I. V. Smirnova, N. Kibiryeva, E. Vidoni, R. Bunag, and L.Stehno-Bittel, “Abnormal EKG stress test in rats with type 1diabetes is deterred with low-intensity exercise programme,”Acta Diabetologica, vol. 43, no. 3, pp. 66–74, 2006.

[13] C. H. Shao, X. H. T. Wehrens, T. A. Wyatt et al., “Exercisetraining during diabetes attenuates cardiac ryanodine receptordysregulation,” Journal of Applied Physiology, vol. 106, no. 4,pp. 1280–1292, 2009.

[14] C. H. Hung, Y. W. Chen, D. Z. Shao, C. N. Chang, Y. Y. Tsai,and J. T. Cheng, “Exercise pretraining attenuates endotoxin-induced hemodynamic alteration in type I diabetic rats,”Applied Physiology, Nutrition and Metabolism, vol. 33, no. 5,pp. 976–983, 2008.

[15] K. R. Bidasee, H. Zheng, C.-H. Shao, S. K. Parbhu, G. J.Rozanski, and K. P. Patel, “Exercise training initiated afterthe onset of diabetes preserves myocardial function: effects onexpression of β-adrenoceptors,” Journal of Applied Physiology,vol. 105, no. 3, pp. 907–914, 2008.

[16] E. Heyman, C. Toutain, P. Delamarche et al., “Exercise trainingand cardiovascular risk factors in Type 1 diabetic adolescentgirls,” Pediatric Exercise Science, vol. 19, no. 4, pp. 408–419,2007.

[17] Y. M. Searls, I. V. Smirnova, B. R. Fegley, and L. Stehno-Bittel, “Exercise attenuates diabetes-induced ultrastructuralchanges in rat cardiac tissue,” Medicine and Science in Sportsand Exercise, vol. 36, no. 11, pp. 1863–1870, 2004.

[18] A. Austin, V. Warty, J. Janosky, and S. Arslanian, “Therelationship of physical fitness to lipid and lipoprotein(a)levels in adolescents with IDDM,” Diabetes Care, vol. 16, no.2, pp. 421–425, 1993.

[19] M. S. Faulkner, “Cardiovascular fitness and quality of lifein adolescents with type 1 or type 2 diabetes,” Journal forSpecialists in Pediatric Nursing, vol. 15, no. 4, pp. 307–316,2010.

[20] S. F. Michaliszyn and M. S. Faulkner, “Physical activity andsedentary behavior in adolescents with type 1 diabetes,”Research in Nursing & Health, vol. 33, no. 5, pp. 441–449, 2010.

[21] R. D’hooge, T. Hellinckx, C. Van Laethem et al., “Influenceof combined aerobic and resistance training on metaboliccontrol, cardiovascular fitness and quality of life in adolescentswith type 1 diabetes: a randomized controlled trial,” ClinicalRehabilitation, vol. 25, no. 4, pp. 349–359, 2011.

[22] S. Zeqiri, A. Ylli, and N. Zeqiri, “The effect of physical activityin glycemia in patients with diabetes mellitus,” MedicinskiArhiv, vol. 61, no. 3, pp. 146–149, 2007.

[23] M. A. Salem, M. A. Aboelasrar, N. S. Elbarbary, R. A.Elhilaly, and Y. M. Refaat, “Is exercise a therapeutic toolfor improvement of cardiovascular risk factors in adolescentswith type 1 diabetes mellitus? A randomised controlled trial,”Diabetology and Metabolic Syndrome, vol. 2, no. 1, article 47,2010.


[24] A. Herbst, R. Bachran, T. Kapellen, and R. W. Holl, “Effectsof regular physical activity on control of glycemia in pediatricpatients with type 1 diabetes mellitus,” Archives of Pediatricsand Adolescent Medicine, vol. 160, no. 6, pp. 573–577, 2006.

[25] K. J. Nadeau, J. G. Regensteiner, T. A. Bauer et al., “Insulinresistance in adolescents with type 1 diabetes and itsrelationship to cardiovascular function,” Journal of ClinicalEndocrinology and Metabolism, vol. 95, no. 2, pp. 513–521,2010.

[26] C. J. Mann, E. Ayuso, X. M. Anguela, and F. Bosch, “Skeletalmuscle metabolism in the pathology and treatment of type 1diabetes,” Current Pharmaceutical Design, vol. 16, no. 8, pp.1002–1020, 2010.

[27] F. C. Howarth, F. M. A. Marzouqi, A. M. S. Al Saeedi, R.S. Hameed, and E. Adeghate, “The effect of a heavy exerciseprogram on the distribution of pancreatic hormones in thestreptozotocin-induced diabetic rat,” Journal of the Pancreas,vol. 10, no. 5, pp. 485–491, 2009.

[28] O. Coskun, A. Ocakci, T. Bayraktaroglu, and M. Kanter, “Exer-cise training prevents and protects streptozotocin-inducedoxidative stress and β-cell damage in rat pancreas,” TohokuJournal of Experimental Medicine, vol. 203, no. 3, pp. 145–154,2004.

[29] H.-H. Huang, L. Novikova, S. J. Williams, I. V. Smirnova, andL. Stehno-Bittel, “Low insulin content of large islet populationis present in situ and in isolated islets,” Islets, vol. 3, no. 1, pp.6–13, 2011.

[30] R. R. MacGregor, S. J. Williams, P. Y. Tong, K. Kover, W. V.Moore, and L. Stehno-Bittel, “Small rat islets are superiorto large islets in in vitro function and in transplantationoutcomes,” American Journal of Physiology, vol. 290, no. 5, pp.E771–E779, 2006.

[31] S. J. Williams, H. H. Huang, K. Kover et al., “Reduction of dif-fusion barriers in isolated rat islets improves survival, but notinsulin secretion or transplantation outcome,” Organogenesis,vol. 6, no. 2, pp. 115–124, 2010.

[32] S. J. Williams, Q. Wang, R. R. MacGregor, T. J. Siahaan, L.Stehno-Bittel, and C. Berkland, “Adhesion of pancreatic betacells to biopolymer films,” Biopolymers, vol. 91, no. 8, pp. 676–685, 2009.

[33] M. Srinivasan, C. S. Choi, P. Ghoshal et al., “β-cell-specific pyruvate dehydrogenase deficiency impairs glucose-stimulated insulin secretion,” American Journal of Physiology,vol. 299, no. 6, pp. E910–E917, 2010.

[34] V. C. Calegari, C. C. Zoppi, L. F. Rezende, L. R. Silveira,E. M. Carneiro, and A. C. Boschero, “Endurance trainingactivates AMP-activated protein kinase, increases expressionof uncoupling protein 2 and reduces insulin secretion from ratpancreatic islets,” Journal of Endocrinology, vol. 208, no. 3, pp.257–264, 2011.

[35] M. Nakata, K. Manaka, S. Yamamoto, M. Mori, and T. Yada,“Nesfatin-1 enhances glucose-induced insulin secretion bypromoting Ca2+ influx through L-type channels in mouse isletβ-cells,” Endocrine Journal, vol. 58, no. 4, pp. 305–313, 2011.

[36] C. A. Machado De Oliveira, M. F. Paiva, C. A. Soares Motaet al., “Exercise at anaerobic threshold intensity and insulinsecretion by isolated pancreatic islets of rats,” Islets, vol. 2, no.4, pp. 240–246, 2010.

[37] V. Shivaswamy, M. McClure, J. Passer et al., “Hyperglycemiainduced by tacrolimus and sirolimus is reversible in normalsprague-dawley rats,” Endocrine, vol. 37, no. 3, pp. 489–496,2010.

[38] K. Robertson, P. Adolfsson, G. Scheiner, R. Hanas, and M. C.Riddell, “Exercise in children and adolescents with diabetes,”Pediatric Diabetes, vol. 10, no. 12, pp. 154–168, 2009.

[39] G. M. Reaven and F. Chang, “Effect of exercise-training on themetabolic manifestations of streptozotocin-induced diabetesin the rat,” Diabetologia, vol. 21, no. 4, pp. 415–417, 1981.

[40] M. A. Kiraly, H. E. Bates, J. T. Y. Yue et al., “Attenuation oftype 2 diabetes mellitus in the male Zucker diabetic fatty rat:the effects of stress and non-volitional exercise,” Metabolism,vol. 56, no. 6, pp. 732–744, 2007.

[41] E. P. Reaven and G. M. Reaven, “Structure and functionchanges in the endocrine pancreas of aging rats with referenceto the modulating effects of exercise and caloric restriction,”Journal of Clinical Investigation, vol. 68, no. 1, pp. 75–84, 1981.

[42] E. Teixeira de Lemos, F. Reis, S. Baptista et al., “Exercisetraining decreases proinflammatory profile in Zucker diabetic(type 2) fatty rats,” Nutrition, vol. 25, no. 3, pp. 330–339, 2009.

[43] M. Colombo, S. Gregersen, M. Kruhoeffer et al., “Preventionof hyperglycemia in Zucker diabetic fatty rats by exercisetraining: effects on gene expression in insulin-sensitive tissuesdetermined by high-density oligonucleotide microarray anal-ysis,” Metabolism, vol. 54, no. 12, pp. 1571–1581, 2005.

[44] M. Maris, G. B. Ferreira, W. D’Hertog et al., “High glucoseinduces dysfunction in insulin secretory cells by differ-ent pathways: a proteomic approach,” Journal of ProteomeResearch, vol. 9, no. 12, pp. 6274–6287, 2010.

[45] Z. Q. Hou, H. L. Li, L. Gao, L. Pan, J. J. Zhao, and G. W.Li, “Involvement of chronic stresses in rat islet and INS-1 cellglucotoxicity induced by intermittent high glucose,” Molecularand Cellular Endocrinology, vol. 291, no. 1-2, pp. 71–78, 2008.

Documents

ExerciseIncreasesInsulinContentandBasalSecretionin ......improvements in exercise-induced glycemic control in T1D mice may be due to enhancement of insulin content and secretion in